Modeled Cost Differences Associated With Use of Levonorgestrel Intrauterine Devices
This analytic model assessed the per-woman cost differences between the use of 2 levonorgestrel-releasing intrauterine devices, from a US payer perspective.
Objectives: To evaluate the differences in mean costs per woman of the use of 2 levonorgestrel-releasing intrauterine devices (IUDs)—Mirena (LNG-M) and Liletta (LNG-L)—from a US payer perspective.
Study Design: A decision analytics model.
Methods: Total healthcare costs associated with IUDs included device cost, costs of insertion and removal, and costs of pregnancy-related outcomes. Pregnancy event rates and costs were obtained from published literature and IUD prescribing information. Total costs of IUDs at 3, 5, and 10 years of use were estimated. A 1-way sensitivity analysis was conducted, as was a Monte Carlo simulation, in which the impact of model parameter variations were evaluated.
Results: At 5 years of contraception use, the mean costs per woman were estimated to be lower for LNG-M than LNG-L ($1089 vs $1614). After 3 and 10 years, the differences in total costs of use of LNG-M and LNG-L were estimated at $69 and —$1160 per woman, respectively. The 1-way sensitivity analysis showed that the device costs of LNG-M and LNG-L have the most impact on the cost differences. The Monte Carlo simulations showed, within a hypothetical cohort of 10,000 women with randomly distributed contraception durations, that approximately 70% are estimated to have cost savings with use of LNG-M versus LNG-L at a mean cost difference of –$335 per woman.
Conclusions: Compared with LNG-L, the use of LNG-M was associated with a slightly higher cost at 3 years, but a cost savings of more than $500 per woman at 5 years and more than $1000 at 10 years.
Am J Pharm Benefits. 2017;9(1):12-16
According to data from the 2006 to 2008 National Survey of Family Growth (NSFG), 49% of pregnancies in the United States were unintended among women aged 15 to 44 years, and more than half occurred among women who were currently using contraception.1 Unintended pregnancies represent a large cost burden to society, with direct medical costs estimated at $4.6 billion annually, with 53% of this cost attributed to imperfect contraceptive adherence.2 Other evidence suggests that the majority of unintended pregnancies among women using contraception result from incorrect or inconsistent use rather than method failure.3 According to the American College of Obstetrics and Gynecology, long-acting reversible contraception (LARC) methods, which include intrauterine devices (IUDs) and implants, are considered the most effective reversible contraceptive methods.4 However, for the years 2011 to 2013, the NSFG reported that LARC methods were used by only 7.2% of women aged 15 to 44 years in the United States.5 The CDC has recommended that providers use a tiered approach to contraceptive counseling, such that the most effective methods are discussed first before presenting information on less effective methods.6
In contrast to short-acting contraception (SARC) methods (eg, oral contraceptives), once in place, IUDs are indicated for use from 3 to 10 years; therefore, efficacy does not rely on user adherence.7 Currently, there are two 52-mg levonorgestrel (LNG)-releasing IUDs—Mirena (LNG-M) and Liletta (LNG-L)—approved by the FDA. LNG-M is approved for intrauterine contraception for up to 5 years and for the treatment of heavy menstrual bleeding in women who choose to use intrauterine contraception as their method of contraception.8 LNG-L is approved for intrauterine contraception for up to 3 years.9 Both products are similar in efficacy for prevention of pregnancy.10-12 As the indicated durations of use differ between these 2 products, this study evaluated their cost differences from a US payer perspective.
Decision Analytics Model Inputs
A decision analytics model was developed in Excel (eAppendix, available at www.ajpb.com). The device costs of the IUDs were based on the Wholesale Acquisition Costs obtained via Red Book (Micromedex Healthcare Series [Thomson Reuters (Healthcare) Inc, Montvale, New Jersey]). Total healthcare costs associated with use of each IUD included the costs of device insertion and removal procedures, as well as costs of pregnancy-related outcomes, and were obtained from the literature (Tables 1 and 2).2,8,9,13,14 Pregnancy-related event rates and costs were obtained from IUD prescribing information and published literature (Table 2).2,8,9,13,14 All costs were inflation adjusted to 2014 cost levels by using the medical care component for the Consumer Price Index.
Total costs of use of IUDs in 2014 US$ were based on device cost and costs of insertion and removal. The costs for IUD insertion and removal used in this study in 2014 US$ are similar to those previously used by Trussell et al in an economic modeling analysis of the average annual costs of available reversible contraceptive methods.15 Costs of pregnancy-related outcomes were obtained from published literature and manufacturer-supplied research data. It was assumed that women used the IUD for the entirety of approved contraceptive duration; additionally, if a woman became pregnant, the IUD was removed, following which, the woman no longer participated in the study cohort. The time of IUD removal procedure or the pregnancy—whichever occurred first—was evaluated at the end of each year of the model calculation and used as the IUD removal time. Additional model assumptions for sensitivity analyses are described below.
Estimate of Cost Differences
Total costs of use of IUDs at 3, 5, and 10 years of contraception use were estimated. Three and 5 years are the indicated durations of contraception of the 2 IUDs, and a timeframe of 10 years was used to further evaluate costs of use of IUDs when women use them for a longer duration. It was assumed that the IUDs were only removed at the end of the respective contraception indication time intervals (3 years for LNG-L and 5 years for LNG-M) or at the time of pregnancy, whichever was earlier. Two alternative scenarios with different distributions of duration of use per device were also evaluated: one with a distribution of use of 30% of women for 3 years, 40% of women for 5 years, and 30% of women for 10 years, which mimicked a relatively even distribution across the timeframes with a small peak at 5 years, and the other with a distribution of use of 50% of women for 3 years, 50% of women for 5 years, and 0% of women for 10 years, which had equal distribution for the indicated contraception durations for the respective IUDs.
Univariate (1-way) sensitivity analyses were conducted to determine the effects of varying the change of a single model input on the medical cost differences associated with the use of LNG-M versus LNG-L at 3, 5, and 10 years. Pregnancy-related outcome rates and costs associated with each IUD, including device cost and costs of device insertion and removal procedures, were varied ± 30%. Additionally, Monte Carlo simulations were used to further assess costs of use of LNG-M versus LNG-L. In each cycle of the Monte Carlo simulation, a random duration of contraceptive use, in whole years (integers), was selected from a Poisson distribution with lambda set to 5 years, and the total medical cost differences of use of LNG-M versus LNG-L were evaluated for each cycle. This process was repeated 10,000 times to complete the Monte Carlo simulation. Descriptive statistics of the total medical cost differences were measured from the results of the 10,000 random Monte Carlo cycles. The 95% confidence intervals (CIs) of the estimated medical cost differences were evaluated as the range between the 2.5 and 97.5 percentiles of medical costs from the Monte Carlo simulation for LNG-M versus LNG-L. Another sensitivity analysis was conducted to assess the costs of the use of LNG-M versus LNG-L, where each cycle of the Monte Carlo simulation used a duration of contraceptive use that was randomly selected from a Gaussian distribution, with a mean of 5 years and standard deviation of 2 years. In this random sampling with Gaussian distribution, when the randomly generated duration was less than 1 year, it was reset to a minimum of 1 year.
At 5 years of contraception use, the total mean costs per woman were lower for LNG-M than for LNG-L ($1089 vs $1614, respectively), with costs of devices and insertion estimated at $911 versus $1449, costs of removal at $125 for both, and costs of pregnancy outcomes at $53 versus $40 (Table 3). At 3 and 10 years, the difference in total costs of the use of LNG-M versus LNG-L were estimated at $69 and —$1160 per woman, respectively (Table 3).
In the alternative scenario, in which 30% of women used either LNG-M or LNG-L for 3 years, 40% for 5 years, and 30% for 10 years, the total mean costs of use of LNG-M and LNG-L were estimated at $1370 and $1907 per woman, respectively, resulting in a difference of —$537 favoring LNG-M (Table 4). In the alternative scenario, in which 50% of women used either LNG-M or LNG-L for 3 years and 50% for 5 years, the total mean costs of use of LNG-M and LNG-L were estimated at $1016 and $1244 per woman, respectively, resulting in a difference of —$228 favoring LNG-M (Table 4). The breakeven point with equal total costs between LNG-M versus LNG-L occurred when approximately 88% of women use the IUD for 3 years and 12% use it for 5 years. This indicates that LNG-M use would provide cost savings as long as more than 12% of women use the IUD for 5 years.
Results of the 1-way sensitivity analyses showed that variations in the device costs of LNG-M and LNG-L had the most impact on the cost difference estimates of use of LNG-M versus LNG-L at 3, 5, and 10 years (eAppendix). Variations in other model inputs, including device insertion and removal costs, pregnancy rates, and costs of pregnancy-related outcomes, did not have major impacts on the cost-difference estimates (Figure). The Monte Carlo simulations showed that within a hypothetical cohort of 10,000 women with randomly distributed contraception durations, approximately 70% are estimated to have cost savings with the use of LNG-M versus LNG-L, with a mean cost difference of —$335 per woman (median = –$457; 95% CI, –$1160 to $258). The additional sensitivity analysis, in which Monte Carlo simulations using a Gaussian distribution were used to generate the random IUD use duration, showed that within a hypothetical cohort of 10,000 women with randomly distributed contraception durations, approximately 74% are estimated to have cost savings with the use of LNG-M versus LNG-L, with a mean cost difference of –$357 per woman (median = –$457; 95% CI, –$1160 to $258). These results are consistent with the default Monte Carlo analysis.
In this cost analysis, the use of LNG-M versus LNG-L was associated with a cost savings of more than $500 per woman at 5 years and more than $1000 at 10 years, but with a slightly higher cost at 3 years. To a US payer, the initial costs of LNG-M and LNG-L are higher than those of SARC methods; however, an analysis of the costs of 16 different contraceptive methods over a 5-year period reported that LNG IUDs are one of the top 3 least expensive methods of contraception.13 Another recent economic analysis that compared the costs of no method, 4 SARC methods (ie, oral contraceptive, ring, patch, and injection), and 3 LARC methods (ie, implant, copper IUD, and LNG IUD) estimated that the 2 least expensive methods were the copper IUD ($304 per woman per year) and the LNG IUD ($308), and that a minimum of 2.1 years of LARC usage would result in cost savings compared with SARC usage.15 This latter study, and others, have also shown that LARC methods become increasingly cost-saving with longer duration of use.2,13,15,16
Currently, LNG-L is approved for intrauterine contraception for up to 3 years in the United States.9 With an approved indication of use of 5 years, LNG-M—based on our analytic model—is associated with lower medical costs versus LNG-L at 5 and 10 years of use. When the durations of use of these LNG IUDs were varied in 10,000 random Monte Carlo simulation cycles, approximately 70% were estimated to have a cost savings with use of LNG-M versus LNG-L. These results suggest LNG-M is a lower-cost option for a US payer under the majority of circumstances.
The Affordable Care Act (ACA), a health insurance reform legislation signed into law on March 23, 2010, in the United States, mandated that health plans must cover FDA-approved contraceptive methods, sterilization procedures, and patient education and counseling for all women with reproductive capacity without patient cost sharing (ie, co-payment, coinsurance, or deductible).17 Thus, knowledge of the medical costs associated with use of different methods of contraception is important to US payers. A recently published study, which evaluated the savings in out-of-pocket expenses for contraceptives among women (n = 790,895) using a national commercial insurer database, reported a 68% decline in out-of-pocket expenses between 2012 and 2013 for IUDs.18 Nevertheless, some women still have out-of-pocket expenses for IUDs.18
Another study conducted by the Kaiser Family Foundation reported that although most of the 20 carriers from the 5 states that were included in the study are in compliance with the ACA provision of access to all FDA-approved contraceptives without cost sharing, some plans still impose limitations.19 In regard to IUDs, 10 of the 20 insurance plans in the Kaiser study covered both LNG IUDs available at the time of the study without cost sharing or limitations, whereas 3 plans covered only 1 of the hormonal LNG IUDs, but not both.19 (The other plans had limitations with IUD coverage and so are not included.) It is important for women to have the option to choose the most appropriate contraceptive method on an individual basis and to have access to the most efficacious and cost-effective contraceptive options. It will be valuable for future studies to examine whether the ACA contraception provision results in an increase in use of LARC versus SARC methods and fewer unintended pregnancies.
This analytic model has some limitations, including that the validity of the assumptions of the model may not be true in real-world settings. Also, the costs of the IUDs, associated procedures, and pregnancy-related outcomes may vary among different payers and treatment settings and are influenced by many other factors. One of the model assumptions was that women were completely compliant using either IUD and thus used the IUD for the approved contraceptive duration. If a woman becomes pregnant, it was assumed that the IUD would be removed and the woman would no longer participate in the study cohort after removal. In the real-world setting, the use of IUDs may differ from this assumption among women who become pregnant. Furthermore, as LGN-L is relatively new to the market, there are no data on the actual duration of its use in the real-world setting; there are limited data for the duration of use of LNG-M. Thus, in our analytic model, we considered different scenarios of IUD duration and conducted sensitivity analyses.
In this cost analysis, use of LNG-M versus LNG-L was associated with a cost savings of more than $500 per woman at 5 years and more than $1000 at 10 years, but had a slightly higher cost at 3 years. Based on this cost analysis, with the limitation in the duration of use for LNG-L, LNG-M represents a lower-cost contraception option than LNG-L.
Author Affiliations: Bayer HealthCare Pharmaceuticals (AL, MM, RL), Whippany, NJ; Novosys Health (ML-S, JL), Green Brook, NJ.
Source of Funding: This study was supported by Bayer HealthCare Pharmaceuticals, Inc.
Author Disclosures: Drs Law, McCoy, and Lynen are employees of Bayer HealthCare Pharmaceuticals, Inc. Drs Lingohr-Smith and Lin are employees of Novosys Health and served as paid consultants to Bayer HealthCare Pharmaceuticals, Inc, for the development of this study and manuscript.
Authorship Information: Concept and design (AL, MM, JL, RL); acquisition of data (AL, MM, RL); analysis and interpretation of data (AL, MM, ML-S, JL, RL); drafting of the manuscript (ML-S, JL); critical revision of the manuscript for important intellectual content (AL, MM, ML-S, JL, RL); statistical analysis (AL, MM, JL); provision of patients or study materials (AL, MM, ML-S, JL, RL); obtaining funding (AL, MM, RL); administrative, technical, or logistic support (ML-S, JL); and supervision (AL, MM).
Address Correspondence to: Amy Law, PharmD, Bayer HealthCare Pharmaceuticals, 100 Bayer Blvd, Whippany, NJ 07981. E-mail: email@example.com.
- Finer LB, Zolna MR. Unintended pregnancy in the United States: incidence and disparities, 2006. Contraception. 2011;84(5):478-485. doi: 10.1016/j.contraception.2011.07.013.
- Trussell J, Henry N, Hassan F, Prezioso A, Law A, Filonenko A. Burden of unintended pregnancy in the United States: potential savings with increased use of long-acting reversible contraception. Contraception. 2013;87(2):154-161. doi: 10.1016/j.contraception.2012.07.016.
- Moreau C, Cleland K, Trussell J. Contraceptive discontinuation attributed to method dissatisfaction in the United States. Contraception. 2007;76(4):267-272.
- American College of Obstetricians and Gynecologists. Long-acting reversible contraception: implants and intrauterine devices. practice bulletin no. 121. Obstet Gynecol. 2011;118(1):184-196. doi: 10.1097/AOG.0b013e318227f05e.
- Daniels K, Daugherty J, Jones J. Current contraceptive status among women aged 15-44: United States, 2011-2013. NCHS Data Brief. 2014;(173):1-8.
- Klein DA, Arnold JJ, Reese ES. Provision of contraception: key recommendations from the CDC. Am Fam Physician. 2015;91(9):625-633.
- Trussell J. Contraceptive failure in the United States. Contraception. 2011;83(5):397-404. doi: 10.1016/j.contraception.2011.01.021.
- Mirena [prescribing information]. Whippany, NJ: Bayer HealthCare Pharmaceuticals, Inc; 2014.
- Liletta [prescribing information]. Parsippany, NJ: Actavis Pharma, Inc; 2015.
- Sivin I, Stern J, Diaz J, et al. Two years of intrauterine contraception with levonorgestrel and with copper: a randomized comparison of the TCu 380Ag and levonorgestrel 20 mcg/day devices. Contraception. 1987;35(3)245-255.
- Luukkainen T, Allonen H, Haukkamaa M, et al. Effective contraception with the levonorgestrel-releasing intrauterine device: 12-month report of a European multicenter study. Contraception. 1987;36(2):169-179.
- Eisenberg DL, Schreiber CA, Turok DK, Teal SB, Westhoff CL, Creinin MD; ACCESS IUS Investigators. Three-year efficacy and safety of a new 52-mg levonorgestrel-releasing intrauterine system. Contraception. 2015;92(1):10-16. doi: 10.1016/j.contraception.2015.04.006.
- Trussell J, Lalla AM, Doan QV, Reyes E, Pinto L, Gricar J. Cost effectiveness of contraceptives in the United States. Contraception. 2009;79(1):5-14. doi: 10.1016/j.contraception.2008.08.003.
- Law A, McCoy M, Lynen R, et al. The prevalence of complications and healthcare costs during pregnancy. J Med Econ. 2015;18(7):533-541. doi: 10.3111/13696998.2015.1016229.
- Trussell J, Hassan F, Lowin J, Law A, Filonenko A. Achieving cost-neutrality with long-acting reversible contraceptive methods. Contraception. 2015;91(1):49-56. doi: 10.1016/j.contraception.2014.08.011.
- Mavranezouli I; LARC Guideline Development Group. The cost-effectiveness of long-acting reversible contraceptive methods in the UK: analysis based on a decision-analytic model developed for a National Institute for Health and Clinical Excellence (NICE) clinical practice guideline. Hum Reprod. 2008;23(6):1338-1345. doi: 10.1093/humrep/den091.
- Women’s Preventative Services Guidelines. Health Resources & Services Administration website. https://www.hrsa.gov/womensguidelines2016/index.html. Accessed January 2017.
- Becker NV, Polsky D. Women saw large decrease in out-of-pocket spending for contraceptives after ACA mandate removed cost sharing. Health Aff (Millwood). 2015;34(7):1204-1211. doi: 10.1377/hlthaff.2015.0127.
- Sobel L, Salganicoff A, Kurani N, et al. Coverage of contraceptive services: a review of health insurance plans in five states. Kaiser Family Foundation website. http://files.kff.org/attachment/report-coverage-of-contraceptive-services-a-review-of-health-insurance-plans-in-five-states. Published April 2015. Accessed August 12, 2015.